The link between agriculture and rural food security in the ecoregions of Mexico: path diagrams and underlying dataset.

In this research, we build two food systems datasets in Mexico; The first one describes the structure of agricultural production units and the second one describes food security aspects of the rural population in these agricultural production units. We also build a third dataset, consisting of path diagrams and path coefficients (derived from Structural Equation Modeling) that relate the first dataset to the second dataset in the four most populated ecoregions of Mexico. The description of the path models and the insights they bring to the current state of food security in Mexican rural households are detailed in an associated article entitled "Is food security primarily associated with smallholder agriculture or with commercial agriculture?: An approach to the case of Mexico using structural equation modeling" (https://doi.org/10.1016/j.agsy.2021.103091). The agricultural variables (in the first dataset) include farm size, destination of the farmer's production, cultivation practice / water management, predominant source of income of the household, land tenure type, crop diversity, agricultural surface expansion, and the presence of forest cover. They are based on the primary data of the full, latest available agricultural census in Mexico and corresponding official land use / land cover data. The second dataset consists of four food security indicators designed and built for the first food security model in Mexico that incorporates food availability, food accessibility and food utilization aspects. They include the Food Self-sufficiency Index (the balance between food production and food consumption), the Food Access Index (inversely related to marginalization), the Entitlement to Public Health Care index, and the Undernutrition Infrequency index (related to hospital sickness records). We provide the path tables and diagrams that describe the links between the agricultural structure and food security. These diagrams provide the first nationwide statistical evidence for the prominent role of smallholder agriculture in rural food security at the national level and at ecoregion scale for a country of the global South. In order to further investigate the structure of the agricultural production units and their relationships with socio-economic, territorial and landscape data, artificial intelligence (i.e. data mining and machine learning) techniques could be performed on this compendium of datasets. The food security data may stir the development of more food security models in Mexico in relation to other drivers such as consumption habits and non-agricultural activities of rural households.